Two-factor authentication for automated teller machine using visual cryptography

Abstract

Automated Teller Machine (ATM) is commonly used nowadays because of its convenience and efficiency. However, it is subjected to several attacks such as password guessing attacks and over-the-shoulder attacks. Attackers pose a threat to account authenticity and financial loss by copying magnetic information of ATM cards. Two-factor authentication is proposed against the fraud. Nevertheless, many two-factor authentication solutions nowadays require extra layer as a barrier, for example, a physical token. These solutions have weaknesses like high cost and difficulty to be implemented. This project aims to provide a simple system model without extra computer devices to perform two-factor authentication. A second authentication is added to the ATM system using visual cryptography proposed by Naor and Shamir. Every user of the new system should have a special ATM card, which is printed with a decoding mask (a transparent scrambled image). When ATM monitor displays an encrypted image in the second authentication, user should overlay his card onto it and input password decrypted for verification. This solution has advantages such as high adaptability, low cost, low weight and convenience. Another function “Dynamic Keypad” is also proposed in this project. A dynamic keypad is presented for PIN input in first authentication. It helps to remedy security fraud of nearby camera and over-the-shoulder attacks. The function is optional for user’s preference.